For Brülosophy exBEERiments, we rely on the triangle test where tasters are presented with three samples of beer, two being identical while one is different. Each taster is then tasked with attempting to identify the odd beer out. We’ve conducted this test hundreds of times, but what’s the rationale behind it? Marshall Schott, founder of Brülosophy, joins us to explain, and we also go over data looking at the influence one’s experience has on their ability to accurately select the unique sample on triangle tests.
4 thoughts on “The Brülosophy Show: How The Brülosophy Triangle Test Really Works”
As your video points out, P represents the probability that the differences that were perceived occurred by chance. I don’t know for a fact, but it seems to me that commercial breweries use the criteria of p<0.05 so that they can be virtually certain (at least 95% probability) that if they make a process change, it will be beneficial. On the other hand, home brewers sometimes don’t have a big investment involved in making a process change. Therefore, different criteria would be appropriate for home brewers. In fact, I think the criteria should be variable, depending on the investment necessary to make the change, possible repercussions if the change is actually harmful to the product, and other things. When I read exbeeriment results, I look primarily at the P value and think about how it relates to my own brewing process and equipment.
Myth Busters. Love it.
One thing that should be investigated is any potential sequence effect with regard to which beers are tasted. For example, if you get the odd beer first, then the two similar, are you more likely to detect the difference compared to when you get the odd beer last after the two similar? Has this been controlled for in all prior xbmts or is it just whatever the researcher happens to pour and line up for each participant one by one? Cheers!
As your video points out, P represents the probability that the differences that were perceived occurred by chance. I don’t know for a fact, but it seems to me that commercial breweries use the criteria of p<0.05 so that they can be virtually certain (at least 95% probability) that if they make a process change, it will be beneficial. On the other hand, home brewers sometimes don’t have a big investment involved in making a process change. Therefore, different criteria would be appropriate for home brewers. In fact, I think the criteria should be variable, depending on the investment necessary to make the change, possible repercussions if the change is actually harmful to the product, and other things. When I read exbeeriment results, I look primarily at the P value and think about how it relates to my own brewing process and equipment.